I agree it is tricky.
The general design of the question_display_options was to divide up all the things that might, or might not, be displayed into a number of categorties that are as independent as possible. This was to enable, in system which supportes multiple question types and behaviour (and indeed potentially different uses of questions like quiz and embedded quetsitions, ...) a way for teachers to express there preferences about what students can see at any time - which can be pretty important for getting the teaching model you want.
However, as you say, specific feedback / marks / correctness cannot ever truly be independent. That is ineherent. So, as a question type creator, you just have to do your best to take all the things which might, or might not, be displayed, and allocate them to one of the categories.
It seeme like Carroll has given us a use case where putting some things into the correctness category might be desirable.
The general design of the question_display_options was to divide up all the things that might, or might not, be displayed into a number of categorties that are as independent as possible. This was to enable, in system which supportes multiple question types and behaviour (and indeed potentially different uses of questions like quiz and embedded quetsitions, ...) a way for teachers to express there preferences about what students can see at any time - which can be pretty important for getting the teaching model you want.
However, as you say, specific feedback / marks / correctness cannot ever truly be independent. That is ineherent. So, as a question type creator, you just have to do your best to take all the things which might, or might not, be displayed, and allocate them to one of the categories.
It seeme like Carroll has given us a use case where putting some things into the correctness category might be desirable.